Am Do., 28. Mai 2020 um 19:40 Uhr schrieb John Cowan <xxxxxx@ccil.org>:

> The future is longer than the past (at least we hope so), but it's the past that controls the present, and the present is a lot of unadapted or minimally adapted copies of SRFI 1.  Of course, one hopes that the predicate is usually an equivalence relation, in which case the bug does not matter.  But when it isn't, the user will read the SRFI, try to use the procedure, fail, and complain.  Changing the SRFI with a warning heads off a lot of complaints.

Adding a warning, updating the sample implementation and asking
maintainers of active Scheme implementations to check their code (this
can be a joint effort by every one of us) seems like a good idea.

This issue won't be the last one where not 100% bugfree (impossible,
I'd say) sample implementations of (other) SRFIs have been used.